You must be new around here. You should read the ER CC transcripts say from Q1 2015 to get a better sense of Musk speak.
Maybe read Ashley Vance's
book or at least read the
excerpt or if not, just this one paragraph as a sample data point:
The proposed timeline for upending the aerospace industry was comically short. One of the earliest SpaceX presentations promised the first complete engine by May 2003, a second engine in June, the body of the rocket in July, and everything assembled by August. A launchpad would be ready by September, and the first launch would take place in November 2003, or about 15 months after the company started. A trip to Mars was naturally slated for somewhere near the end of the decade. “Elon has always been optimistic,” said Kevin Brogan, an early SpaceX recruit. “That’s the nice word. He can be a downright liar about when things need to get done. He will pick the most aggressive time schedule imaginable assuming everything goes right, and then accelerate it by assuming that everyone can work harder.”
You are correct that Elon is sometimes aggressive. I think the book excerpt is spot on. I have said many times here that I think Elon makes the most optimistic possible schedule the official goal way too often. So if something should take 24 months, he says it should be 12, his team does it in 18 and it is branded a late failure. I wish he would stop doing it but his insane style probably shaved off 6 months in my example. Its a good result but bad optics.
As for Tesla's supposedly horrible track record:
The roadster: Late. They didn't know what they were doing and thought they could re-use much more from Lotus than they could.
The Model S: On time. The first car they did from the ground up, and a mission critical one. Designed to be great, not cheap.
The Model X: Famously late. Designed to be loopy complicated, to shore up high-end demand until the 3 could be built. Since the model S was proving to be more popular than expected, the X was not needed financially so it was allowed to slack. Wasn't critical.
The Model 3: I predict will be similar to the model S. They will produce a few "on time" then ramp quickly. Designed to be simple to manufacture, the BIW line will take fewer steps, be more idiot proof and the interior will be designed to be assembled and installed by robots. The drive train is already simple. The final assembly will still be done partially with people, but the steps will be fewer and throughput high.
Here is another way to think of it. Making EV's SHOULD be very simple. A large part of the machining cost of an ICE is simply not present in an EV so that cost would emerge if all things were equal. But we haven't see that case yet.
The Roadster wasn't simple since it wasn't their design. Basically hand-built custom cars.
The Model S wasn't simple because it didn't need to be. Wildly profitable at 100k+ it just needed to be good.
The Model X wasn't simple because Elon doubled down on fancy instead of making it simpler. Same philosophy though, wildly profitable at 100k.
The Model 3 will be the first instance of nothing unnecessarily physically fancy. True, fancy sensors and hardware is being installed but that is just to future-proof and self-driving reduces to a software problem that they will solve later. But even with a HUD, I think they are considering the interior to be a single dashboard unit that can be assembled elsewhere and installed in a snap.
I imagine the interior being offloaded to the underused upstairs or other facilities for 2017. So it isn't hard to see how they will get to market and ramping pretty quickly.